COVID-۱۹ diagnosis model using lung CT-scans based on Deep Learning

Due to the epidemic of COVID۱۹ and the diagnosis of this disease based on CT scans, it is very important to accurately diagnose sick people with healthy people. Therefore, the purpose of this paper is to build a model that can detect the health or infection of the lungs using CT scan of healthy people and people with the disease . In this study, deep learning framework on Kaggle was used to diagnose lung health or infection using CT scans of healthy individuals and people with the disease COVID۱۹. The data used were collected by referring to the official media, the general. Python programming language version ۳.۷ is used to implement models.The number of images of people with this disease is ۳۴۹ and the number of images of lungs of healthy people is ۳۹۷. The images used to teach this model are ۲۹۹ x ۲۹۹ x ۳, with ۴۲۵ images for training, ۱۱۸ for validation, and ۲۰۳ for network testing. The network were retrained by lungs CT images. After ۱۵ rounds of training, the accuracy level was ۰.۹۴. The precision, recall, F۱of this model are up to ۰.۹ that has the highest ones in comparison to other works.

کلیدواژه ها:

Coronavirus disease (COVID-۱۹) ، Deep learning framework ، Kaggle ، CT scan